Arcee CTO says Chinese open-weight AI models not inherently dangerous
8 reported
As Chinese open-weight AI models grow in capability and popularity, debate has intensified over whether they should be banned. Lucas Atkins, CTO of U.S. open-source AI lab Arcee, argues that Chinese models are no more dangerous than any other open-source software a company might use. Atkins stated that there is no way for a model’s creator to access or control a model once it is run in a user’s own environment. He noted that large organizations can put any model through security testing and post-training processes before use. Atkins also said that the conversation should focus on fostering a good open ecosystem in the U.S. rather than banning Chinese models. Arcee itself benefits from Chinese models because they are open, allowing the startup to learn from them and build on top of them. Atkins expressed respect for the researchers building those models and said the way to compete is to release a better model.
What’s reported
Arcee is a U.S. open-source AI lab building open models as a homegrown alternative to Chinese models.
Lucas Atkins, Arcee’s CTO, said Chinese open-weight models are not inherently dangerous.
Atkins stated that there is no way for a model creator to access or control a model once it is run in a user’s own environment.
Large organizations can put any model through security testing and post-training processes.
Atkins said the conversation should be about fostering a good open ecosystem in the U.S., not banning Chinese models.
Arcee benefits from Chinese models because they are open, allowing the startup to learn from and build on them.
Atkins expressed respect for the individual researchers building Chinese models.
Atkins said the way to compete with Chinese models is to release a better model.
Key figures
Lucas Atkins, CTO of Arcee
Sources: TechCrunch
